G × E Interaction and stability analysis to identify resistant mungbean (Vigna radiata L.) genotypes against Boeremia leaf spot using AMMI and GGE biplot analysis

Abstract

Boeremia Leaf Spot (BLS), caused by Boeremia exigua, threatens mungbean cultivation worldwide, exacerbated by its wide host range, multiple pathogenic strains, and environmental influences. A thorough understanding of the host-pathogen-environment interactions is critical for effective disease management. To identify mungbean genotypes resistant to BLS, we conducted multi-environment trials. Initially, 70 genotypes were screened under controlled conditions, and 10 were selected for field trials across four environments over three years. GGE biplot analysis revealed the importance of considering both genetic and environmental factors when evaluating resistance. Based on average disease severity (DS) scores across locations and years, two genotypes, (G5) AG-11 (DS = 4.1) and G7 (AK-58) (DS = 3.4), which demonstrated consistent resistance to BLS across multiple test environments. These genotypes not only exhibited strong disease resistance but also showed stable adaptation both in case of disease resistance and yield output [ (G5) AG-11 (Y = 9.02qtls/ha) and (G7) (AK-58) (y = 12.15 qtls/ha) and therefore were identified as promising resistant candidates. The Additive Main Effects and Multiplicative Interaction (AMMI) model, in conjunction with the GGE biplot, effectively analyzed genotype-environment interactions and identified optimal evaluation sites and “mega-environments” for BLS resistance. These findings provide valuable insights for future breeding programs focused on integrating resistance traits into new mungbean varieties.

Clinical trial study

Not applicable.

Introduction

Mungbean (Vigna radiata), a member of the Fabaceae family, is an important warm-season legume crop with a diploid chromosome number of 2n = 22 [1]. It is believed to have originated in the Indian subcontinent and later spread to various tropical and subtropical regions of Asia, and Africa [2]. Mungbean is well-suited for cultivation in arid and semi-arid regions due to its short life cycle, drought tolerance, and ability to thrive under high temperatures, making it a valuable crop for sustainable agriculture in water-limited environments [3]. It is primarily grown for its protein-rich seeds, which serve as a staple in many diets, particularly in Asia, and are also used in sprouts, flour, and animal feed. Additionally, mungbean plays a vital role in soil fertility improvement through biological nitrogen fixation [4]. Despite its adaptability, mungbean production faces significant economic losses due to both biotic and abiotic stresses [5]. In biotic stresses, insect pests such as pod borers (Maruca vitrata), whiteflies (Bemisia tabaci), and aphids (Aphis craccivora) pose major threats. Moreover, several fungal, bacterial, and viral pathogens also affect mungbean, leading to devastating yield losses [6]. The yield loss in mungbean ranges from 11 to 50% at national level and upto 70% in other crops at global level [7]. Among major fungal diseases, Boeremia Leaf Spot (Boeremia exigua) is newly emerging, devastating disease that further threatens mungbean productivity. Boeremia exigua is a fungal pathogen belonging to the family Didymellaceae [8]. It is a complex species group with multiple host specific variants, causing diseases in a wide range of crops. Boeremia exigua primarily causes leaf spots, stem cankers, and seedling blights in various crops, including legumes like mungbean. It produces visible circular to irregular brown lesions with dark margins, often coalescing under high humidity (above 80%) and sunken, elongated lesions on stems and petioles, which may crack and lead to plant lodging. Dark, sunken spots on pods, seed discoloration, and infected seeds may appear dark or shriveled, reducing their market value and viability [9]. Boeremia exigua can be morphologically differentiated from other species in the Boeremia genus by number of traits, such as pycnidia which are dark brown to black, flask shaped structures embedded in host tissues [10]. Conidia are hyaline to light brown, oval to cylindrical, measuring about 3–6 × 1–2 μm. Mycelium is septate, branched, and dark colored in culture. It grows on PDA (potato dextrose agar) with a dark brown to grayish appearance, sometimes producing aerial mycelium.

Mungbean, a billion-dollar crop essential for food security in India, China, and Myanmar, faces significant threats from Boeremia exigua and other pathogens [11]. Identifying resistant genotypes through robust selection strategies is crucial for sustainable production. The genotype × environment (G × E) interaction is a fundamental consideration in mungbean breeding, particularly for developing cultivars suited to diverse and challenging conditions such as arid and semi-arid regions [12]. G × E interaction occurs when different genotypes respond variably across environments, meaning a variety performing well in one location may not perform similarly in another due to differences in climate, soil, water availability, or biotic stresses [13]. Understanding these interactions is essential to improve selection accuracy, identify broadly adapted or specifically adapted genotypes, and enhance breeding efficiency [14]. In regions prone to abiotic stresses such as drought, salinity, and heat, G × E analysis facilitates the identification of stress-tolerant and resilient genotypes. It also helps uncover genetic mechanisms underlying environmental responses, enabling targeted breeding. Moreover, insights from G × E interactions inform the design of breeding strategies, including the delineation of breeding zones, implementation of multi-environment trials (METs), and application of genomic selection models to enhance prediction accuracy.

One of the primary goals in mungbean improvement is yield stability across environments. G × E analysis enables the selection of genotypes with consistent performance and low sensitivity to environmental variation, thereby reducing production risks [15]. It also supports the development of varieties optimized for resource-use efficiency, which is particularly valuable in low-input farming systems typical of dryland. Location-specific cultivars guided by G × E data can maximize land, water, and nutrient use efficiency, promoting sustainable agriculture [16]. Genotype-by-environment (G × E) interaction studies, using models like the Additive Main Effects and Multiplicative Interaction (AMMI) model and the Genotype + Genotype by Environment (GGE) biplot, help evaluate mungbean performance across diverse conditions [17]. The AMMI model analyzes genotype-environment interactions, while the GGE biplot visualizes stability and adaptability, aiding in multi-location and multi-year trials [18]. These tools assist breeders in developing disease-resistant, high-yielding varieties adaptable to varying climates. Variations in B. exigua inoculum across environments complicate genotype rankings, making the identification of optimal testing sites with high repeatability essential. Clustering similar environments into “mega environments” enhances screening efficiency and ensures consistent evaluations [19]. However, defining mega-environments depends on repeatable GEI patterns, which can change over time due to climate shifts or pathogen evolution. Recent studies highlight the effectiveness of G × E analysis in breeding for disease resistance. As global demand for mungbean increases, particularly in water-limited and degraded soils, such research is vital to reducing economic losses from fungal diseases [20]. However, while genotype × environment interaction (G × E) studies have significantly contributed to breeding efforts targeting yield and abiotic stress tolerance in mungbean, limited research has focused on biotic stress particularly fungal diseases such as Boeremia leaf spot (BLS) in the context of G × E interaction. The current literature lacks comprehensive multi-environment studies that use robust models like AMMI and GGE biplot to analyze genotype performance under disease pressure. Furthermore, there is a notable gap in the evaluation of test environments based on their discriminatory ability and representativeness for disease resistance screening. Without this knowledge, breeding programs risk inefficient site selection and suboptimal resource allocation. Moreover, the absence of clearly defined mega-environments hampers the development of region-specific, disease-resistant cultivars tailored to varying agro-ecological zones. Addressing these gaps is critical for advancing disease-resilient mungbean varieties and enhancing sustainable production, particularly in arid and semi-arid regions prone to BLS outbreaks. Despite the economic importance of mungbean and the growing threat posed by Boeremia leaf spot disease, limited research has been conducted to evaluate genotype performance and stability under disease pressure across diverse environments. Most previous studies have focused on yield traits or abiotic stress tolerance, with insufficient attention to biotic stress-related G × E interactions, especially using robust statistical models like AMMI and GGE biplot analysis. Moreover, there is a lack of information on the discriminatory power and representativeness of different testing sites, which is crucial for efficient genotype evaluation and resource allocation in breeding programs. The absence of well-defined mega-environment classifications further limits the development of targeted, region-specific disease-resistant mungbean cultivars. This study addresses these gaps by integrating multi-environment trial data with AMMI and GGE biplot analysis to assess both disease resistance stability and genotype performance across diverse locations. By identifying stable, high-yielding mungbean genotypes resistant to Boeremia leaf spot and determining optimal testing sites based on statistical indices (discrimination ability, representativeness, and desirability), the study provides a scientific basis for site selection, germplasm screening, and targeted breeding strategies. The classification of locations into distinct mega-environments offers a strategic framework to enhance the efficiency of mungbean improvement programs under disease stress, ultimately contributing to sustainable legume production in disease-prone agro-ecological zones.

In this study, we employed the GGE biplot approach using multi-location data to assess GEI for BLS resistance. Our objectives included identifying stable, high-performing mungbean genotypes and determining optimal testing sites based on discrimination power, representativeness, and desirability index. Additionally, we categorized testing locations into distinct “mega environments” to guide future breeding and screening efforts effectively.

Materials and methods

Initial testing

A thorough evaluation of 70 mungbean genotypes (list of all genotypes along with their pedigree and geographic origin are given in Table S1), encompassing released varieties, was performed in 2022–2023 to determine their resistance to Boeremia Leaf Spot (BLS), as outlined in Table S2. After conducting a preliminary screening, a selection of 10 resistant mungbean genotypes was made for further testing in multiplication and multi-year trials.

Fig. 1

Symptoms of Boeremia Leaf Spot on infected mungbean plants. (a) Infected mungbean leaf. (b)Boeremia exigua culture (c) Conidia of Boeremia exigua

Screening under in-vitro conditions

Boeremia exigua was grown on potato dextrose agar (PDA) in a BOD incubator. A spore suspension of 1 × 10⁶ conidia per mL was prepared by flooding the plates with sterile distilled water, scraping the surface, filtering through sterile cheesecloth of 12 micron pore size, and adjusting the concentration using a hemocytometer. Thirty-day-old mungbean plants were sprayed with this spore suspension using a handheld atomizer set at a pressure of approximately 20 psi, producing fine droplets with an average size of 100–150 microns. This droplet size ensured uniform coverage and effective inoculation of the leaf surface while minimizing runoff, and to ensure adequate moisture, the plants were covered with clear polyethylene bags for a duration of 72 h. A control plant was maintained without inoculation under the same conditions. After 72 h, the bags were removed, and the plants were moved to a greenhouse (with temperature 25 ± 2 °C with a relative humidity of 80–90%, conditions conducive to fungal development and disease progression) for three weeks for further observations. Symptoms were recorded at regular intervals, and a disease rating scale ranging from 1 to 9, based on Iqbal [21], was used to assess disease severity. To confirm the pathogen’s identity in accordance with Koch’s postulates, it was re-isolated from the symptomatic leaves, and the new cultures were compared to the original ones [22].

Multi environment testing

A comprehensive evaluation was performed on 16 elite mungbean lines, including the susceptible check G8 (SK-86), to assess their vulnerability to Boeremia Leaf Spot (BLS) across four distinct locations over three consecutive growing seasons (2022–2024) under natural epiphytotic conditions. The trial was conducted at FoA Wadura Sopore, SKUAST-K Shalimar, KVK Kupwara, and MRCFC Khudwani. FoA Wadura Sopore and MRCFC Khudwani were specifically identified for ongoing testing over the three years due to the consistent natural occurrence of BLS, establishing them as disease hotspots. Detailed information regarding the plant materials used in this study is provided in Table 1. The genotypes were cultivated following standard agricultural practices in a randomized complete block design (RCBD) with three (3) replications and suitable plant spacing. Inoculum pressure was enhanced by interspersing a row of the susceptible genotype (WMB-1) after every third row of the test genotypes, surrounding the test block. Additionally, meteorological data for all experimental sites is presented in Table 2.

Screening and disease documenting

BLS intensity was rigorously analyzed once the susceptible check reached 50% disease severity, with each genotype subsequently evaluated for infection following Bakr protocol [23]. Since BLS typically appears post flowering in mungbean leaves from 10 randomly selected plants were evaluated for disease severity during the pod filling stage at 60 days (Fig. 1). A disease rating scale, adapted from Iqbal [21], was used where 1, which indicates no observable symptoms, 9 represents more than 70.1% of the affected foliage. Intermediate categories were defined, with 2 representing 0.1–10% foliage affected, 3 for 10.1–20%, and continuing up to 8 for 60.1–70% foliage affected. Annual data on disease intensity from all experimental sites was compiled for each genotype to perform a GGE biplot analysis.

AMMI Model

The Additive Main Effects and Multiplicative Interaction (AMMI) model was employed to analyze genotype × environment interaction (GEI) and assess the stability and resistance of mungbean genotypes against Boeremia Leaf Spot (BLS) across multiple test environments. The AMMI approach combines analysis of variance (ANOVA) to partition the additive effects of genotypes and environments, with principal component analysis (PCA) to capture the multiplicative interaction effects (GEI). This model effectively identifies both high-performing and stable genotypes across varying agro-climatic conditions. The AMMI model can be mathematically represented as:

$$Yij = \mu + {G_{\rm{i}}} + Ej + \mathop \sum \limits_{k = 1}^n {\lambda _k}{{\rm{\alpha }}_{ik}}{{\rm{\gamma }}_{jk}} + {{\rm{\rho }}_{ij}}$$

where Y_ij is the observed response (e.g., disease severity or yield) of genotype (i) in environment (j); (µ) is the grand mean; (Gi) and (E_j) are the additive effects of genotype and environment, respectively; ($\:{\uplambda\:}$_k) is the singular value for the (k^th) principal component axis (PCA); (α_ik) and (γ_jk) are the genotype and environment eigenvectors for axis (k); and (ρ_ij​) is the residual error. The model was implemented to identify genotypes with consistent BLS resistance and to explore GEI patterns, thereby facilitating the selection of stable, disease-resistant genotypes suited for diverse growing conditions.

Construction of GGE biplot

The work of Yan et al. [24] expanded the application of the GGE biplot for genotypes evaluation and mega-environments identification in multi-environment trial (MET) data. They focused on the genotype main effect (G) and genotype-environment interaction (GEI), minimizing the influence of minor environmental main effects (E). The GGE biplot is constructed by plotting genotype and environment scores on the first principal component (PC1) against the second principal component (PC2), both derived through singular value decomposition (SVD) of environment-centered data according to the provided equation [25].

$$Yij = \mu + {e_j} + \mathop \sum \limits_{n = 1}^N {\lambda _n}{\gamma _{in}}{{\rm{\delta }}_{jn}} + {\varepsilon _{ij}}$$

Where:

Yij is the mean severity of ith genotype (i = 1,….,i) in the j^th environment (j = 1,…,j).

µ is the grand mean.

N is the number of principal components retained in the model.

$\:{\:e}_{j}$is environment deviations from grand mean.

λ_n is the eigenvalue associated with principal component analysis axis.

γ _in is the genotype PC score for axis.

δ_jn is the environment PC score for axis.

ϵ_ij is the residual error term.

Table 1 Plant material used in the study

The analysis of the MLT dataset regarding the response of mungbean genotypes to BLS severity was conducted without scaling (scaling = 0), aiming to create an environment-centered G × E table (Centering = 2). The assessment of genotypes relied on genotype-based single value portioning (SVP = 1), while environmental evaluations utilized environment-based single value portioning (SVP = 2) [26]. The optimal genotype was identified based on its mean performance and stability across various environments [27]. Genotype evaluations were performed and visualized using the “average environmental coordination” (AEC) perspective of the GGE biplot, facilitating the comparison of mean disease scores and stability across environments within a mega-environment [28]. The suitability of experimental sites was assessed using the “discriminatory power and representativeness perspective” of the GGE biplot, where discriminatory power was determined by the magnitude of the environmental vector, and representativeness was indicated by the angle formed with the “Average Environment Coordinate/Axis” (AEC/AEA) [23]. Furthermore, genotype performance across diverse experimental conditions and the classification of test environments into distinct “mega-environments” were determined using the GGE biplots’ “which won where” approach [29].

Table 2 Climate factors at experimental sites concerning days with rain, overall precipitation, and temperature throughout growth period of Mungbean

Data analysis

An assessment of the effects of environments, genotypes, and their interactions on BLS was conducted using analyses of variance (ANOVA) alongside mixed model analyses and GGE biplots, with the “METAN” package in R software version 4.3.3.

Results

Genotypic response towards BLS infestation

Across multiple tested locations, mungbean genotypes exhibited varying levels of susceptibility to Boeremia leaf spot (BLS). The ANOVA conducted on BLS infection indicated that genotype, environment, and their interaction (GEI) had significant effects (Table 3). The analysis of variation sources showed that the environment accounted for 49.12% of the overall variation, while the genotype × environment interaction contributed 14.11%. This underscores the significant influence of environmental factors on BLS severity among genotypes across diverse experimental sites. The data presented in Fig. 2 further highlights the varying success rates of mungbean genotypes and the significant differences in BLS severity observed across locations. Kupwara (KP-1) showed the highest variability in the first year (2022-23), followed by Shalimar (SL-1). In contrast, Anantnag (AG-1) exhibited the least variability, followed by Wadura (W-1), highlighting the strong genotype-environment interaction in host-pathogen dynamics. In the second and third years, the frequency of resistant genotypes increased in Kupwara during the second and third years. Conversely, Anantnag (AG-1) had the highest frequency of susceptible genotypes in the first year. Wadura recorded the highest BLS severity at 8.2 in third year W-3, followed closely by Anantnag, which had a severity of 7.8 in two environments (AG-1, AG-3) during the first and third years, and 7.9 in AG-2 in the second year. Wadura also exhibited notable severity in the first (W-1) and second years (W-2). Conversely, Kupwara had the lowest severity, with values of 1.7 and 1.8 in the first (KP-1) and second years (KP-2), and 2.1 in the third year (KP-3). Severity readings in Shalimar were 3.0 in the first year (SL-1), 3.2 in the second year (SL-2), and 2.5 in the third year (SL-3). The susceptible check (WMB-1) showed BLS severity between 6.6 and 8.2, with an overall average of 7.4, reflecting a consistent disease pressure across various years and locations. Over the years, the genotypes AK-58, AG-11, and SM-8 exhibited moderate resistance, whereas WB-13, KP-108, and SM-14 were classified as moderately susceptible. The performance of these genotypes varied significantly across different locations and years, as illustrated in Fig. 3. Based on average disease severity (DS) scores across locations and years, AK-58, AG-11, and SM-8 were identified as moderately resistant, each with a mean disease rating below 5. Specifically, their average DS scores were 3.4 for AK-58, 4.1 for AG-11, and 5.1 for SM-8.

Fig. 2

Heatmap visualization of BLS severity in mungbean across four locations. The x-axis shows the tested environments. Locations are denoted as W– (Wadura), AG- (Anantnag), K- (Kupwara), SL- (Shalimar) for year 1 (2022): W1, AG1, KP1, SL1; year 2 (2023): W2, AG2, KP2, SL2; year 3 (2024): W3, AG3, KP3, SL3. The y-axis shows the tested genotypes. The plot legend for disease severity depicts the 1–9 scale for BLS severity rating in color. BLS stands for Boeremia Leaf Spot”

Evaluation of genotypes based on mean performance and stability

The biplot effectively demonstrates the ranking of genotypes according to their average performance across different locations and throughout the year, employing the “average environmental coordination” (AEC) axis to illustrate mean performance and stability (Fig. 3). The analysis reveals that PC1 and PC2 contribute 90.79% and 7.78% of the total variation in BLS scoring, respectively, when factoring in environmental influences. The AEC axis, depicted as a single arrow extending from the center of the biplot, indicates a greater severity of BLS in the genotypes. Upon examining the figure, it is clear that G1(SM-8), G5 (AG-11), and G7 (AK-58) are located to the right of the biplot origin, suggesting they experience lower BLS infestation levels. Conversely, the susceptible check G8 (SK-86), along with G2 (WB-13), G4 (KP-108), G3 (SM-16), G6 (SM-14), G9 (BL-22), and G10 (SK-23), are positioned to the right of the origin, indicating a higher disease severity.

Fig. 3

GGE biplot illustrating mean versus stability of 10 mungbean genotypes across four experimental sites. The data were not transformed (transform = 0) and centered by the means values of the environments (centering = 2). The biplot was build upon “row metric preserving.” The testing locations were as follows: year 1 (2024) - W1, AG1, KP1, SL1; year 2 (2022) - W2, AG2, KP2, SL2; year 3 (2023) - W3, AG3, KP3, SL3

Table 3 ANoVA for BLS intensity in Mungbean tested at 4 different experimental sites

The stability of genotypes was represented by a line featuring two arrows, referred to as the “AEC ordinate.” The extent of a genotype projection from this line serves as an indicator of its stability; longer projections correlate with reduced stability. Ideally, genotypes should exhibit low disease intensity and minimal projection on the AEC axis. In this context, AG-11 (G5) emerges as an “ideal” genotype, showcasing moderate resistance and high stability, characterized by a short projection along the AEC axis. Genotypes that are closer to this ideal genotype are considered more “desirable.” For instance, AK-58 (G7) was identified as a “desirable” genotype due to its moderately susceptible nature and consistent performance across various locations. Conversely, genotype SM-8 (G1) displayed a moderately susceptible response but had greater projections from the AEC axis, indicating variability in its reaction to BLS across different locations.

Assessment of testing locations: discriminative vs. Representativeness and desirability index

The interrelation among the experimental sites is illustrated through the environmental vector feature of the GGE biplot, where each site is connected by a line that denotes its environmental vector. The cosine of the angle between two environmental vectors indicates their level of association. A significant relationship was found between Kupwara (across all three years) and Shalimar (across two years), as well as between Wadura and Anantnag throughout all three years. Additionally, Shalimar in the third year (SL-3) showed a close relationship with Wadura and Anantnag across all years. These sites exhibited acute angles, which signify a consistent genotypic response to BLS intensity.

Fig. 4

The perspective of delimitation versus representativeness of experimental sites was compared using a GGE biplot of 10 mungbean genotypes evaluated at four experimental sites. The data were not transformed (transform = 0), and centered using mean values of the environments (centering = 2). The biplot was created using “row metric preserving”. The locations are: For year 1 (2022): W1, AG1, KP1, SL1; Year 2 (2023) W2, AG2, KP2, SL2; Year 3 (2024) W3, AG3, KP3, SL3

In the context of the GGE biplot framework, three essential variables are vital for the evaluation of test locations: “discrimination power,” “representativeness,” and the “desirability index.” The “discriminating ability” is defined by the length of the environmental vector that corresponds to the standard deviation at each experimental site. Over the years, Kupwara and Shalimar have consistently exhibited the longest environmental vectors, whereas Wadura has recorded the shortest projection (Fig. 4). Consequently, Shalimar and Kupwara have been classified as “discriminating locations” due to their capacity to effectively differentiate among genotypes.

The concept of “representativeness” in experimental sites is determined by the acute angle formed between the environmental vectors and the “AEC abscissa.” Throughout the years, Shalimar and Wadura have exhibited acute angles with the “AEC abscissa,” marking them as the “most representative” sites. The “desirability index” of these experimental locations merges their “discriminative” power with their “representativeness.” Consequently, Shalimar, boasting the highest desirability index, is identified as the “ideal” site for BLS screening, while Wadura is classified as a “supplementary” site.

Identification of mega environments and “ which won where”

A polygonal diagram that indicates “which won where” is constructed by interleaving the genotypes to a limited extent from the origin. Subsequently, vertical lines known as “ equality lines” are drawn from the origin to both sides of the polygon, dividing it into various sectors. The interaction between genotypes and environments is analyzed using “symmetric scaling,” which incorporates singular value partitioning to highlight both genotype and environment. This visualization is instrumental in identifying significant environments and genotypes with specific resilience traits.

Fig. 5

“Which-won-where” perspective of GGE biplot used to analyze 10 mungbean genotypes across 4 experimental sites. Data was not transformed (transform = 0), and were centered by means of the environments (centering = 2). The biplot was based on “row metric preserving”. Locations are: For year 1 (2022): W1, AG1, KP1, SL1; year 2 (2023) W2, AG2, KP2, SL2; year 3 (2024) W3, AG3, KP3, SL3

Genotypes situated at the vertices of the polygon typically exhibited the strongest associations between genotype and environment, highlighting either the highest or lowest performers within that specific area. In this research, the polygon is divided into four distinct sectors that represent two “mega environments (ME),” thereby confirming the presence of crossover interactions (Fig. 5). It is important to note that the test sites of Shalimar, Wadura, Anantnag, and Kupwara have not interacted with one another over the years, resulting in the formation of a single “mega environment.” Over the years, they demonstrated stable genotypic responses to the severity of BLS. The genotype positions at the vertices of each “mega environment,” which included Kupwara, Anantnag, Shalimar, and Wadura, revealed that (G9) and (G8) emerged as the top-performing genotypes, for each mega environment. In MEII, G2 (WB-13) and G10 (SK-23) showcased the most impressive performance.

Discussion

BLS represents a major fungal threat to mungbean crops, with no complete resistant sources identified to date [30]. The identification of BLS-resistant mungbean genotypes is made more challenging by the variability of the pathogen and its wide host range. It is vital to find genotypes that demonstrate low BLS intensity and consistent performance across various environments to effectively incorporate these resistant sources into future breeding programs and management strategies [31]. However, the genotype-by-environment interaction complicates the testing process across multiple environments [32]. Furthermore, the intricate nature of host-pathogen interactions adds another layer of complexity to the selection of durable resistant sources under varying conditions and over time. Assessing various genotypes across different locations presents a considerable challenge. Alongside the need to identify genotypes with lasting resistance, it is crucial to select appropriate testing sites for efficient multi-environment trials. AMMI model combines analysis of variance (ANOVA) for additive main effects (genotype and environment) with principal component analysis (PCA) for the multiplicative interaction term, effectively partitioning variation and providing a robust estimate of genotype stability through metrics such as the AMMI Stability Value (ASV). In contrast, the GGE biplot focuses on genotype and G × E effects only excluding the main effect of environment and is particularly suited for visualizing patterns such as “which-won-where,” ranking genotypes based on performance and stability, and identifying ideal test environments. While AMMI is statistically rigorous and suited for quantifying stability and understanding interaction components, GGE biplot offers a more intuitive and graphical interpretation, especially valuable for mega-environment delineation and test site evaluation. The use of both models in tandem ensures a more holistic and accurate assessment of genotype performance and adaptability, enhancing the precision of selection decisions in resistance breeding. This dual approach strengthens the identification of stable, high-performing genotypes and optimal testing sites, thereby contributing to the development of resilient mungbean cultivars under diverse agro-climatic conditions. The GGE biplot methodology provides a valuable approach to address these issues, enabling the evaluation of genotypes and the identification of optimal experimental locations, which can subsequently be categorized into specific “mega environments,” irrespective of their agroecological zones [33].

This research evaluated 10 mungbean genotypes across four different locations over a span of three years, highlighting a significant genotype × environment interaction in relation to BLS intensity. The analysis of variance (ANOVA) demonstrated that environmental factors exerted the greatest influence on BLS severity, followed by the genotype × environment (G x E) interaction [34]. The performance of the genotypes across various sites exhibited notable differences, indicating the occurrence of crossover interactions. Previous studies have established that fluctuations in environmental conditions at experimental sites can greatly affect BLS severity [35]. Furthermore, variations in genotypic traits and pathogen virulence may also play a role in disease progression [36]. The pronounced G x E interaction emphasizes the importance of conducting multilocation evaluations of genotypes for BLS before establishing final genotype rankings.

The analysis of various locations revealed that Kupwara (KP-2) and Shalimar (SL-2) demonstrated relatively low severity of Boeremia Leaf Spot (BLS), exhibiting responses that ranged from resistant to moderately resistant. In contrast, Anantnag (AG-1, AG-2, and AG-3) and Wadura (W-1 and W-2) showed a range of responses from highly susceptible to susceptible. As a result, Anantnag and Wadura have been classified as “hot spots” for screening mungbean varieties resistant to BLS. The increased severity of BLS observed during the kharif season in Anantnag highlights its aggressive nature. The region experienced a higher number of rainy days in recent years, which contributed to the escalation of disease intensity. The consistent susceptibility of the control variety “SK-86” across different years and locations indicates a sufficient level of infection during the screening process, thereby aiding in the selection of specific genotypes. In multi-location assessments, it is essential to screen genotypes tailored to a mega environment with non-crossover associations [37]. Therefore, when selecting an ideal genotype, both its average performance and stability must be considered [38].

In the context of the GGE biplot analysis, the “AEC abscissa” indicates superior average performance, which reflects the genotype contribution (G), while the “AEC ordinate” signifies genotype stability and illustrates the genotype role in genotype × environment interaction [22]. Among the genotypes evaluated, AG-11 (5) was found very close to the “AEC abscissa,” exhibiting low prediction on the “AEC ordinate” with high stability, thus qualifying as an “ideal” genotype. The categorization of other genotypes is based on their relative position to this ideal genotype in the biplot. Additionally, AG-11 (5) and AK-58 (7) should be significantly incorporated into mungbean resistance breeding programs aimed at combating BLS.

The findings of this study highlighted a strong link between Kupwara and Shalimar, suggesting that genotypic responses to BLS severity are consistent. Given the low BLS severity recorded in Kupwara, it may be omitted from subsequent mungbean genotype assessments. In the context of GGE biplot analysis, an “ideal” experimental location should be selected based on its ability to differentiate genotypes, represent the “mega environment,” and maintain a high “desirability index.”

The term “mega environment” describes a group of similar sites that reliably yield the best genotypic responses throughout the year [39]. It is recommended to choose experimental sites with longer vector lengths for their insightful nature, while sites with sharp angles in relation to the “AEC abscissa” are regarded as representative [40]. The desirability index of a genotype, which combines its discriminatory power with the environment representativeness, is essential for selecting experimental locations to optimize resource use [41]. Anantnag has been identified as a “hot spot” for BLS and is considered the “ideal” site for experiments, with Wadura serving as a “supplementary” location for future mungbean genotype evaluations against BLS. These locations are expected to assist in identifying genotypes with wider adaptability. In contrast, Shalimar and Kupwara have been classified as ineffective for BLS screening, with limited potential to identify superior genotypes.

This research categorized all experimental sites into four distinct “mega environments,” each linked to successful genotypes. This classification underscores the presence of a crossover genotype-by-environment interaction and emphasizes the necessity of breeding for specific adaptability. A “mega environment” is defined as a group of similar locations that demonstrate uniform genotypic responses and consistently support a primary set of genotypes throughout the year [42]. Identifying mega environments is crucial for the assessment of genotypes and experimental sites, facilitating the targeted deployment of specific genotypes in each area [43]. Dividing experimental sites into sub-regions based on their repeatability is crucial for mungbean resistance breeding programs, as it helps target stable environments for selection. In this context, repeatability was measured as the proportion of total phenotypic variance attributed to genotypic variance across environments, indicating the consistency of genotype performance over multiple trials [44]. Consequently, this study offers valuable insights into the most effective mega environments for evaluating mungbean genotypes in relation to BLS. The research identified AG-11 (5) and AK-58 (7) as the most promising genotypes, which may aid in developing resistance to this disease in mungbean.

Conclusion

This study underscores the critical role of understanding genotype-environment interactions (GEI) in identifying mungbean genotypes resistant to Boeremia Leaf Spot (BLS), a significant foliar disease affecting crop health and yield. GEI analysis is essential in resistance breeding, as it helps distinguish genotypes with stable performance across diverse environments from those whose resistance is environment-specific. Through multi-environment trials (METs), the research evaluated a diverse set of mungbean genotypes under varying agro-climatic conditions. Advanced statistical tools such as the GGE biplot and the Additive Main Effects and Multiplicative Interaction (AMMI) model were employed to dissect the GEI and visualize the performance and stability of genotypes. The analysis revealed two standout genotypes, “AG-11 (5)” and “AK-58 (7),” which demonstrated consistent resistance to BLS across multiple test environments. These genotypes not only exhibited strong disease resistance but also showed stable adaptation, making them promising candidates for use in breeding programs. The study highlights the importance of integrating both genetic potential and environmental responses into selection strategies, ensuring that newly developed cultivars maintain resistance under diverse field conditions. The findings contribute valuable insights for mungbean improvement programs by providing a robust framework for selecting resilient genotypes. Ultimately, such an approach supports the development of BLS-resistant mungbean varieties, contributing to enhanced crop resilience, reduced yield losses, and improved sustainability in mungbean cultivation.